Meta’s AI agent has been blocked from using Amazon.com

Access to major e-commerce platforms for AI agents is no longer a given.

Meta's AI agent, Muse, has been blocked from accessing Amazon.com to make purchases. Amazon displayed an error message stating that "Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use." This action prevents Muse from functioning as a shopping agent on the platform. The block is seen as a strategic move by Amazon, which operates its own AI foundation models and inference platform. While Meta's Muse is noted for having a low hallucination rate among AI models, the risk of errors in purchases could lead to Amazon having to manage customer and vendor complaints. This suggests Amazon may be hesitant to fully embrace AI-driven commerce until AI models are more robust, even if they see potential in the market.

Amazon's decision to block Meta's Muse signals a critical juncture for AI agents in commercial applications. Companies are exercising control over how AI interacts with their core business, prioritizing brand safety and operational stability over early adoption of agentic commerce. This creates a fragmented environment where AI agents may not have seamless access to the internet's full functionality, potentially hindering their development and utility for tasks requiring real-world transactions.

The move suggests that AI agents, despite advancements in accuracy, are still perceived as a liability by large platforms wary of errors and the associated customer service fallout. Amazon's stance indicates a preference for a more cautious, phased rollout of AI-driven commerce, possibly after further AI development or the establishment of clear operational frameworks and liability protections. This hesitation from a key player like Amazon could slow down the broader integration of AI agents into everyday online shopping.